Ron,
I thought I would just let you and the list know that I have managed to make some good progress with porting the Geode GX1/cs5530 code to V2.
At this point I have only tested up to getting a kernel booting - it is not yet mounting a root filesystem, but that will come soon.
One of the first things I did was to go through the RAM detection code in V1 with a fine toothcomb, and managed to find a number of bugs in that code, so that now detects every single type of DIMM and SODIMM I have here to test with correctly, so that can be contributed back to the V1 tree. I then re-wrote that code in C and managed to get that running under V2 correctly. The Geode does not have an SMBUS so getting the memory size reported by LinuxBIOS has required a nasty hack. I have put a hook into the northbridge code where it probes the pci bus for mc_dev and slotted in some code return the correct values from the MC_BANK_CFG register of the Geode. This RAM reporting will also apply to devices like the STPC as well, so there should probably be a clean way put together to cater for devices without an SMBUS - I think reporting with a fake smbus routine may be a bit of overkill, because once memory sizing on the STPC is done correctly, it can be reported from a single register as well.
I then created an irq_tables.c file from Linux running from the standard BIOS.
Next step was to get Etherboot running and the elfboot code works like a charm in V2 - I still cannot figure out what I am doing wrong in the V1 code, but that probably does not matter much any more!
I then set up my DHCP server and a TFTP server and first time round managed to get the result shown by the serial console output below - I was expecting the kernel to die because there is in fact no hard drive at /dev/hda3 at the moment, and for this project I do not need any hard disk, as the filesystems will all be embedded into on-board NAND flash.
LinuxBIOS-1.1.8.0Normal Tue Apr 19 12:17:40 CEST 2005 starting... Setting up default parameters for memory Sizing memory Probing for DIMM0 Found DIMM0 Page Size: 00001000 Component Banks: 2 Module Banks: 2 DIMM size: 02000000 Probing for DIMM1 Found DIMM1 Page Size: 00001000 Component Banks: 4 Module Banks: 1 DIMM size: 04000000 MC_BANK_CFG = 14204320 Copying LinuxBIOS to ram. Jumping to LinuxBIOS. LinuxBIOS-1.1.8.0Normal Tue Apr 19 12:17:40 CEST 2005 booting... Enumerating buses... Finding PCI configuration type. PCI: Using configuration type 1 PCI_DOMAIN: 0000 enabled PCI: pci_scan_bus for bus 0 PCI: 00:00.0 [1078/0001] ops PCI: 00:00.0 [1078/0001] enabled PCI: 00:0e.0 [1317/0985] enabled PCI: 00:12.0 [1078/0100] enabled PCI: 00:12.1 [1078/0101] enabled PCI: 00:12.2 [1078/0102] enabled PCI: 00:12.3 [1078/0103] enabled PCI: 00:12.4 [1078/0104] enabled PCI: 00:13.0 [0e11/a0f8] enabled PCI: pci_scan_bus returning with max=00 done Allocating resources... Reading resources... Done reading resources. Setting resources... I would set ram size to 0x18000 Kbytes PCI: 00:0e.0 10 <- [0x0000001000 - 0x00000010ff] io PCI: 00:0e.0 14 <- [0x00febe2000 - 0x00febe23ff] mem PCI: 00:0e.0 30 <- [0x00febc0000 - 0x00febdffff] rom PCI: 00:12.1 10 <- [0x00febe3000 - 0x00febe30ff] mem PCI: 00:12.2 20 <- [0x0000001400 - 0x000000147f] io PCI: 00:12.3 10 <- [0x00febe4000 - 0x00febe407f] mem PCI: 00:12.4 10 <- [0x00febe0000 - 0x00febe0fff] mem PCI: 00:13.0 10 <- [0x00febe1000 - 0x00febe1fff] mem Done setting resources. Done allocating resources. Enabling resourcess... PCI: 00:00.0 cmd <- 147 PCI: 00:0e.0 cmd <- 143 PCI: 00:12.0 cmd <- 14f PCI: 00:12.1 cmd <- 142 PCI: 00:12.2 cmd <- 141 PCI: 00:12.3 cmd <- 142 PCI: 00:12.4 cmd <- 142 PCI: 00:13.0 cmd <- 142 done. Initializing devices... Root Device init PCI: 00:00.0 init Devices initialized Copying IRQ routing tables to 0xf0000...done. Verifing copy of IRQ routing tables at 0xf0000...failed Wrote linuxbios table at: 00000500 - 00000690 checksum f8b7
Welcome to elfboot, the open sourced starter. January 2002, Eric Biederman. Version 1.3
23:stream_init() - rom_stream: 0xfffc0000 - 0xfffcffff Found ELF candiate at offset 0 Loading Etherboot version: 5.2.6 Dropping non PT_LOAD segment New segment addr 0x20000 size 0x12740 offset 0xb0 filesize 0x89fc (cleaned up) New segment addr 0x20000 size 0x12740 offset 0xb0 filesize 0x89fc Loading Segment: addr: 0x0000000000020000 memsz: 0x0000000000012740 filesz: 0x00000000000089fc Clearing Segment: addr: 0x00000000000289fc memsz: 0x0000000000009d44 Jumping to boot code at 0x20000 ROM segment 0x0000 length 0x0000 reloc 0x00020000 CPU 238 Mhz Etherboot 5.2.6 (GPL) http://etherboot.org Tagged ELF for [Tulip] Relocating _text from: [00020000,00033750) to [05eec8b0,05f00000) Boot from (N)etwork or (Q)uit? Probing pci nic... [centaur-p] centaur-p: [chip: ADMtek Comet] rev 17 at 1000 centaur-p: Vendor=1317 Device=0985 centaur-p: 00:0C:41:EB:02:45 at ioaddr 1000 Searching for server (DHCP)... ..Me: 192.168.1.153, Server: 192.168.1.201, Gateway 192.168.1.1 Loading 192.168.1.201:vmlinuz.elf ...(ELF)... ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ ............................................................................ done Firmware type: LinuxBIOS Linux version 2.6.11 (root@monster1) (gcc version 3.3.4 (pre 3.3.5 20040809)) #1 Tue Apr 19 16:09:06 CEST 2005 BIOS-provided physical RAM map: BIOS-e820: 0000000000000000 - 0000000000000700 (reserved) BIOS-e820: 0000000000000700 - 00000000000a0000 (usable) BIOS-e820: 00000000000f0000 - 00000000000f0400 (reserved) BIOS-e820: 0000000000100000 - 0000000006000000 (usable) 96MB LOWMEM available. DMI not present. ACPI: Unable to locate RSDP Allocating PCI resources starting at 06000000 (gap: 06000000:fa000000) Built 1 zonelists Kernel command line: console=ttyS0,38400 root=/dev/hda3 Initializing CPU#0 PID hash table entries: 512 (order: 9, 8192 bytes) Detected 233.869 MHz processor. Using tsc for high-res timesource Console: colour dummy device 80x25 Dentry cache hash table entries: 16384 (order: 4, 65536 bytes) Inode-cache hash table entries: 8192 (order: 3, 32768 bytes) Memory: 94720k/98304k available (1436k kernel code, 3160k reserved, 468k data, 132k init, 0k highmem) Checking if this processor honours the WP bit even in supervisor mode... Ok. Mount-cache hash table entries: 512 (order: 0, 4096 bytes) Working around Cyrix MediaGX virtual DMA bugs. CPU: Cyrix MediaGXtm MMXtm Enhanced stepping 02 Checking 'hlt' instruction... OK. NET: Registered protocol family 16 PCI: Using configuration type 1 ACPI: Subsystem revision 20050211 ACPI: Interpreter disabled. Linux Plug and Play Support v0.97 (c) Adam Belay SCSI subsystem initialized PCI: Probing PCI hardware PCI: Probing PCI hardware (bus 00) audit: initializing netlink socket (disabled) audit(3445976884.071:0): initialized Total HugeTLB memory allocated, 0 Initializing Cryptographic API PCI: Fixup for MediaGX/Geode Slave Disconnect Boundary (0x41=0x9c) isapnp: Scanning for PnP cards... isapnp: No Plug & Play device found Real Time Clock Driver v1.12 i8042: ACPI detection disabled i8042.c: Can't read CTR while initializing i8042. Serial: 8250/16550 driver $Revision: 1.90 $ 48 ports, IRQ sharing enabled ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A io scheduler noop registered io scheduler anticipatory registered io scheduler deadline registered io scheduler cfq registered RAMDISK driver initialized: 16 RAM disks of 64000K size 1024 blocksize loop: loaded (max 8 devices) mice: PS/2 mouse device common for all mice input: PC Speaker NET: Registered protocol family 2 IP: routing cache hash table of 512 buckets, 4Kbytes TCP established hash table entries: 4096 (order: 3, 32768 bytes) TCP bind hash table entries: 4096 (order: 2, 16384 bytes) TCP: Hash tables configured (established 4096 bind 4096) NET: Registered protocol family 1 VFS: Cannot open root device "hda3" or unknown-block(0,0) Please append a correct "root=" boot option Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
I notice the kernel is not doing a PCI probe - I wonder why?
I obviously have a fair amount more cleanup to do, and to make sure I can mount filesystems and that the Interrupts do indeed work. There is another area I may have a look at and that is to attempt to plug in support for the notorious Geode VSA stuff, which at last AMD have released proper specs for.
Hamish -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.17 - Release Date: 19.04.2005
On Tue, Apr 19, 2005 at 05:27:21PM +0200, Hamish Guthrie wrote:
I thought I would just let you and the list know that I have managed to make some good progress with porting the Geode GX1/cs5530 code to V2.
I'll say. Great job!
PCI: Using configuration type 1
[..]
PCI: Probing PCI hardware PCI: Probing PCI hardware (bus 00)
[..]
PCI: Fixup for MediaGX/Geode Slave Disconnect Boundary (0x41=0x9c)
[..]
VFS: Cannot open root device "hda3" or unknown-block(0,0) Please append a correct "root=" boot option Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
I notice the kernel is not doing a PCI probe - I wonder why?
It seems to find the bus according to the (bus 00) message, but no devices on that bus. I seem to remember the Geode PCI had some quirks with regard to addressing, but I could be mistaken.
I obviously have a fair amount more cleanup to do, and to make sure I can mount filesystems and that the Interrupts do indeed work.
The kernel output you sent showed no IDE controller being detected, and hence no hda, but that's just the result of Linux not finding any PCI devices. Interrupts aren't important just yet either, Linux has to see the device first.
There is another area I may have a look at and that is to attempt to plug in support for the notorious Geode VSA stuff, which at last AMD have released proper specs for.
Nice. It would complete the circle for the Geodes. The VSMs are binary-only modules though, so I guess they can cause some licensing problems if linked into LinuxBIOS. :\
//Peter
-
Hamish Guthrie -
Peter Stuge